Poster No:
1120
Submission Type:
Abstract Submission
Authors:
Shiva Sedghi1,2
Institutions:
1Université Laval, QC, Canada, 2CERVO Brain Research Center, QC, Canada
First Author:
Shiva Sedghi
Université Laval|CERVO Brain Research Center
QC, Canada|QC, Canada
Introduction:
Working memory, known as the ability of the human brain to temporarily hold and manipulate sequential items, is at the critical juncture between memory, attention, and perception, and is central to a range of cognitive functions and goal-directed behaviors [3, 4]. Load-dependent increase of brain activity and brain rhythmic activity in the theta frequency range in fronto-parietal networks have been associated with working memory performance [1, 2]. In parallel, recent studies have suggested that internally generated sequences of neural representations can be reactivated or 'replayed' to support memory consolidation and learning. Replay has also been proposed as a mechanism that enables planning and action preparation based on past experiences [5-10]. However, the potential relationship between replay and active working memory retention and manipulation remains unexplored.
Methods:
In this study, we utilized magnetoencephalography (MEG) to measure neural activity before, during after associative learning and during the working memory processes. We recruited 30 young participants in 2 consecutive MEG sessions with pre-learning exposure, associated learning, post-learning evaluation, and working memory tasks. During the working memory task, participants had to mentally recall, and regenerate the previously learned sounds based on given instructions. We used multivariate classification techniques to identify neural patterns associated with memorized items and investigate the dynamics of these patterns during working memory retention and manipulation periods. In addition, we explored the neural mechanisms associated with associative learning in humans by using a combination of behavioral, MEG, and machine learning approaches.
Results:
Our analysis demonstrated sustained theta activity within a distributed network, indicating the temporal cooperation of brain regions during and after learning. ERF and Hilbert transform analysis also revealed a learning-associated beta desynchronization over fronto-temporal and parietal lobes, regions known to be involved in working memory functions. Furthermore, we detected evidence of backward replay during the manipulation and retention phases of working memory processes.
Conclusions:
We observed sustained theta activity in the defined regions of interest after a memory-related task, which may be attributed to the role of theta oscillation in facilitating long-range communication between distant brain regions and can serve as a marker for the formation and retention of memory associations. These findings provide further support for the involvement of theta activity in cognitive processes.
We also observed functional changes in the beta range (12–30 Hz) across several brain regions after the learning task. Notably, these alterations encompassed a decrease in beta power. The desynchronization was observed over a distributed network, which may be related to flexible processing and integration of information between different brain regions. Our results align with previous studies that have shown beta desynchronization during memory tasks and suggest a learning-associated beta desynchronization.
While our study provides valuable insights into the neural mechanisms underlying working memory and proposes that the fast-backward replay may play a crucial role in strengthening memory traces, it is important to note that our analyses on the sequenceness of the backward and forward replay during working memory retention and manipulation are still ongoing. Our findings of fast-backward replay during working memory phases are an exciting development but need to be validated. Further research is needed to elucidate the exact functions and mechanisms of backward replay during the manipulation phase of working memory and to determine the implications of this finding for our understanding of the neural basis of working memory processes.
Higher Cognitive Functions:
Imagery 2
Learning and Memory:
Working Memory 1
Modeling and Analysis Methods:
Classification and Predictive Modeling
Multivariate Approaches
Novel Imaging Acquisition Methods:
MEG
Keywords:
Cognition
Learning
MEG
Memory
Other - Replay
1|2Indicates the priority used for review
Provide references using author date format
Here are the citations arranged in alphabetical order by the first author's last name, with journal names given in full:
1. Albouy, P. (2017). 'Selective entrainment of theta oscillations in the dorsal stream causally enhances auditory working memory performance', Neuron, vol. 94, no. 1, pp. 193-206.
2. Albouy, P. (2018). 'Driving working memory with frequency‐tuned noninvasive brain stimulation', Annals of the New York Academy of Sciences, vol. 1423, no. 1, pp. 126-137.
3. Baddeley, A. (1992). 'Working memory', Science, vol. 255, no. 5044, pp. 556-559.
4. Baddeley, A. (2010). 'Working memory', Current Biology, vol. 20, no. 4, R136-R140.
5. Eldar, E. (2020). 'The roles of online and offline replay in planning', eLife, vol. 9, e56911.
6. Liu, Y. (2019). 'Human replay spontaneously reorganizes experience', Cell, vol. 178, no. 3, pp. 640-652.
7. Liu, Y. (2020). 'Temporally delayed linear modelling (TDLM) measures replay in both animals and humans', eLife, vol. 10.
8. Nour, M. M. (2021). 'Impaired neural replay of inferred relationships in schizophrenia', Cell, vol. 184, no. 16, pp. 4315-4328.
9. Wimmer, G. E. (2020). 'Episodic memory retrieval success is associated with rapid replay of episode content', Nature Neuroscience, vol. 23, no. 8, pp. 1025-1033.
10. Wittkuhn, L. (2021). 'Dynamics of fMRI patterns reflect sub-second activation sequences and reveal replay in human visual cortex'.